Overview
Wei-Yin Loh is a statistician (Professor, University of Wisconsin–Madison) known for classification and regression tree methodology, especially algorithms with unbiased split-variable selection.
Key Contributions / Features
- Unbiased tree algorithms: developer of GUIDE, QUEST, and CRUISE, which select the split variable by a statistical significance test before choosing the split point, avoiding the selection bias of CART's greedy exhaustive search.
- Review of tree methods (Loh 2011): an overview comparing CART-style algorithms on selection bias, missing values, and interaction detection (CART).
Related